ljm42

I finally got USB passthrough to work! Thought I'd document it for anyone else with an ASROCK E3C226D2I motherboard who is trying to boot an unRAID VM on an unRAID host per CHBMB's instructions: https://www.linuxserver.io/index.php/2015/12/14/creating-an-unraid-virtual-machine-to-run-on-an-unraid-host/ The biggest problem was the bios. Everything fell into place once I installed bios 3.30 and BMC 00.22.00 from here: http://www.asrockrack.com/general/productdetail.asp?Model=E3C226D2I#Download Until I did that, it was impossible to isolate the USB controllers - everything appeared on one controller regardless of where it was physically plugged in. I also bought this: http://smile.amazon.com/StarTech-Motherboard-4-Pin-Header-USBMBADAPT/dp/B000IV6S9S and installed it on one of the internal USB headers. My VM's flash drive is plugged in to that, while all the external USB ports go to the host. (I'm actually not sure which controller the built-in internal USB Type A port is connected to, once I got a working config I didn't go back and check that.) This is the modification I made to syslinux to stub out the appropriate USB controller: label unRAID OS menu default kernel /bzimage append pci-stub.ids=8086:8c2d initrd=/bzroot label unRAID OS Safe Mode (no plugins) kernel /bzimage append pci-stub.ids=8086:8c2d initrd=/bzroot unraidsafemode And here is the corresponding xml change: <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x00' slot='0x1a' function='0x0'/> </source> </hostdev> Everything else you need is in CHBMB's blog post. Note: there is still one problem I haven't been able to figure out... when I boot the host, I get this error: An operating system wasn't found. Try disconnecting any drives that don't contain an operating system. Press any key to restart. Once I press a key it goes to the normal unraid boot menu. It seems to be trying to boot off the vm's flash drive, and none of my bios tweaking has managed to prevent that. --- It occurs to me that this is a lot of pain to go through just because unRAID requires the USB drive to have the label "UNRAID". It would be much easier if it could follow this logic: If a USB drive labeled UNRAID_VM exists, and there is no USB drive labeled UNRAID, then mount the UNRAID_VM drive as UNRAID and continue booting. This would allow the host to boot normally, because a USB drive named UNRAID_VM won't cause any issues when one named UNRAID already exists. And the user would simply pass the UNRAID_VM drive to the guest, which would rename it to UNRAID and boot normally.

The only downside to backing up your flash drive is that restoring the super.dat file will trigger a parity check, right? Seems like a small price to pay to have a full backup of all your settings. This is covered in the OP.

Looks like you've run into this bug: https://lime-technology.com/forum/index.php?topic=40826.0

For duplicate file names, it looks for files with the same path and name on different disks. This is the same behavior as how unRAID sees duplicate names. To find duplicate files regardless of their location, look at the duplicate hashes file. Ps. "other content" in your example above denotes two files with the same path/name but different content (their hashes aren't equal). In that case, it is working perfectly. Thanks for this!

Is that type of error logged in nginx's error.log? Because that file is in the config folder under logs/nginx/ and should be easily accessible. When nginx doesn't load, nothing shows up in error.log I looked at firstrun.sh, and it seems like we should be able to evaluate the return values from: service nginx start but I didn't have any luck with that. What if you add nginx -t right before starting the service? That gives some nice debug info, assuming it will be echo'd to the docker log. Ah... I can't keep track of which versions of Linux use apt-get vs yum or something else, so I didn't even try it. There are probably other people like me who would benefit from having it pre-installed. I used an online generator, but it definitely made me nervous.

The LetsEncrypt container is awesome! And thanks a ton for your example configs, they helped a lot. The only real problem I had was that the docker logs give no indication when nginx doesn't start due to a bad config file. Is there any way this could be detected and flagged in the docker log? Also, would you consider adding apache2-utils to the container? Then we could use the htpasswd tool to create .htpasswd files. Nginx and Apache use the same file format. Thanks!

Fixed. If you rename cAdvisor when adding it, the resource monitor (and pop ups) will recognized it Aha! I figured out what was going on... if cadvisor is running, it works as you said (and is really cool!) If cadvisor is stopped then it prompts you to install even if it is already installed. Could CA detect that cadvisor is stopped and offer to start it? I tend to stop the dockers I'm not actively using, which is why I ran into this. Fixed Cool, thanks! This is a problem. Its doable, but ultimately it requires a *major* rewrite of a ton of sections of code to have this happen. It will display in the installed applications section, but as far as CA is concerned, the app that's available on the feed is not installed -> and in a way its not because you customized it. The only places CA displays customized apps is in the installed or previously installed section. Like I said, its a big problem for me to fix, and by far the easiest solution is to ask lsio to name their apps to be something more "natural" No worries, I really don't want to ask you to do major rewrite for this I can rename my Docker if it really starts to bother me.

Thanks, that's much cleaner I moved the files and updated to the 2016.03.30 version Can you clarify what the Find button is looking for? My duplicate_file_names.txt file still just shows this: disk3,disk2 [other content] Movies/todo/test.txt So if it is looking for files at the same path on different disks, then it is working correctly. But if it is looking for duplicate filenames (without considering the path) then it is missing a lot.

A few days ago I clicked on this note to install cAdvisor: Docker Application Resource Monitor For more detailed results, click here (This will install cAdvisor) But today, the same note is displayed. Any chance it could detect that cAdvisor is already installed and offer to run it instead? I have the LS.IO "plex" app, but I named it "PlexMediaServer". When I click the icon in CA, it shows a lot of useful information, but not "% CPU" or "Memory". Would it be possible to show those even though I renamed the app? The information is shown on the actual Resource Monitor page. Also, when I search for Plex-related apps, CA doesn't seem to know that I already have the LS.IO docker installed. I'm sure this is because I renamed it, and if it isn't fixable that is fine, but I thought I'd mention it. BTW, thanks a ton for all of the work you put into CA. It really is amazing. I too would like to see it become a standard part of unRAID, but not until you're done with it

Actually, you can just go to the tools page of this plugin and press the Clear button. This will delete any attributes that were added before the files were excluded, and once the attributes are gone the files will no longer be processed.

Random crashes or disconnects are most likely caused by not giving Crashplan enough RAM. See the Q&A in the OP.

Wow that was fast, thanks bonienl! I pressed "find" and got: Reading and sorting hash files Including... disk1.export.hash Including... disk1.export.20160214.bad.hash Including... disk2.export.hash Including... disk3.export.hash Including... disks.export.20160109.bad.hash Including... disks.export.20160111.bad.hash Including... disks.export.20160115.bad.hash Including... disks.export.20160118.bad.hash Including... disks.export.20160129.bad.hash Including... disks.export.20160130.bad.hash Finding duplicate file names Duplicate file names found! See log file: duplicate_file_names.txt. Even though it says "Duplicate file names found", my duplicate_file_names.txt file just contains: [other content] So I purposefully created two files: /mnt/disk1/Movies/todo/test.txt /mnt/disk2/Movies/todo/test.txt and ran it again. Now duplicate_file_names.txt contains: disk3,disk2 [other content] Movies/todo/test.txt So I'm glad that file shows up, but it seems like the first run should have picked up some too, especially after I saw the data from the next run... I ran "find" a third time with the "include duplicate file hashes" option. It ran for about 15 minutes and created a 243k line file! This is amazing data. I see I need to exclude a few more directories (such as anything SVN related), but it did identify duplicate photos, videos, and other files that I need to clean up. I am definitely surprised by what it dug up. I need to figure out how to parse the file for groups of files (so I can work on all photos at once, etc) but overall this is very cool.

That was the reasoning behind converting the date and time to a sortable numeric format. Could you run ipmi-sel from the command line and see if you have some weird dates or times that don't fall in the format | Mar-27-2016 | 00:57:28 |. oh cool. Yes, I have several that have "PostInit" in both the date and time columns. See attached. Added a couple checks for valid date. Should be fixed now. It works perfectly! And the archive tab is great too, thanks a lot dmacias!

That was the reasoning behind converting the date and time to a sortable numeric format. Could you run ipmi-sel from the command line and see if you have some weird dates or times that don't fall in the format | Mar-27-2016 | 00:57:28 |. oh cool. Yes, I have several that have "PostInit" in both the date and time columns. See attached. impi.txt

Hey bonienl, I was looking at these scripts: https://lime-technology.com/forum/index.php?topic=35183.0 https://lime-technology.com/forum/index.php?topic=7018.msg68073;topicseen#msg68073 and I realized both could efficiently be replaced by this plugin if it did some parsing of the export files: It could look for files with the same name (not including the path). This would identify issues with a file existing at both /mnt/disk1/share/file and /mnt/disk2/share/file, for instance. The hashes wouldn't necessarily be the same. It could look for files with identical hashes, regardless of file name, to identify true duplicates. I think this would be a nice way to get extra value out of these export files, if you are up for it

Take your time, and thanks for all of your work on this plugin! I still have 2016.03.14 installed on one of my servers, and I noticed another issue - the sort by timestamp column does an alphabetical sort rather than a date sort. Sorry The archive function sounds great!

Hi dmacias, I updated the IPMI plugin from 2016.03.14 to 2016.03.25 on unRAID 6.1.9, and now my Event Log doesn't load. It just displays "Please wait, retrieving event information..." I did some digging and found that there is a problem with ipmi_events.php. This url: http://[server ip]/plugins/ipmi/include/ipmi_events.php returns: Fatal error: Call to a member function format() on boolean in /usr/local/emhttp/plugins/ipmi/include/ipmi_helpers.php on line 70 Is there anything else I can provide to help diagnose this? Aside from that I see you added an Archived Events tab, very cool! So to use it, just turn the switch on and hit "Clear All"?

You can get the md5 hash with this command: getfattr -n user.md5 /path/to/filename If you are using something other than md5, do this first to see what attributes are available: getfattr /path/to/filename

Just another data point for you... this weekend I precleared a 4TB drive using the beta plugin, then added it to the array and ran a parity check. No errors! Using the 2016.03.15 version of the plugin on unRAID 6.1.9.

People started reporting problems transcoding to RAM with Plex 0.9.14: https://lime-technology.com/forum/index.php?topic=40937.msg419495#msg419495

I know it's a little ambiguous. Maybe I could label the switch Backup on Clear or Save on Clear and the lable the 'dellete all' 'clear all' With backup on it sets the ipmi-sel option to --post-clear instead of just --clear. This clears the SEL after displaying all the SEL records. The output is then stored in a gzip like this ipmi_event_log-2016-03-12-214431.gz in /boot/config/plugins/ipmi/logs. I'm thinking we don't even need a switch, just have it backup automatically when you press delete all? What sort of format is the backup in? Is it something the plugin can use somehow or is it something else?

This is great dlandon, thanks!

Everything is working great! The only thing I'm not sure of is the backup. I set the switch to "backup on", do I need to click "remove all" to have it do something? What actually happens if I do that?

This is amazing gfjardim, thank you

Thanks for this dmacias! It looks like it is working well on 6.1.9 I like that there are two temps and two fans in the footer. The temps don't show the units © though, is that to save space? Is it important to show the ID in the basic view? Not sure how I would use that in day-to-day operation The event log loads *much* faster now, wow! Would it make sense to reverse the default sorting of the event log, so that the newest is on top? Is it possible to test notifications without sticking your finger in a fan? The instructions reference ipmitool, which is no longer installed Thanks again!